Nutrient availability, especially of nitrogen (N) and phosphorus (P), is of major importance for every organism and at a larger scale for ecosystem functioning and productivity. Changes in nutrient availability and potential stoichiometric imbalance due to anthropogenic nitrogen deposition might lead to nutrient deficiency or alter ecosystem functioning in various ways. In this study, we present 6 years (2014–2020) of flux‐, plant‐, and remote sensing data from a large‐scale nutrient manipulation experiment conducted in a Mediterranean savanna‐type ecosystem with an emphasis on the effects of N and P treatments on ecosystem‐scale water‐use efficiency (WUE) and related mechanisms. Two plots were fertilized with N (NT, 16.9 Ha) and N + P (NPT, 21.5 Ha), and a third unfertilized plot served as a control (CT). Fertilization had a strong impact on leaf nutrient stoichiometry only within the herbaceous layer with increased leaf N in both fertilized treatments and increased leaf P in NPT. Following fertilization, WUE in NT and NPT increased during the peak of growing season. While gross primary productivity similarly increased in NT and NPT, transpiration and surface conductance increased more in NT than in NPT. The results show that the NPT plot with higher nutrient availability, but more balanced N:P leaf stoichiometry had the highest WUE. On average, higher N availability resulted in a 40% increased leaf area index (LAI) in both fertilized treatments in the spring. Increased LAI reduced aerodynamic conductance and thus evaporation at both fertilized plots in the spring. Despite reduced evaporation, annual evapotranspiration increased by 10% (48.6 ± 28.3 kg H2O m−2), in the NT plot, while NPT remained similar to CT (−1%, −6.7 ± 12.2 kgH2O m−2). Potential causes for increased transpiration at NT could be increased root biomass and thus higher water uptake or rhizosphere priming to increase P‐mobilization through microbes. The annual net ecosystem exchange shifted from a carbon source in CT (75.0 ± 20.6 gC m−2) to carbon‐neutral in both fertilized treatments [−7.0 ± 18.5 gC m−2 (NT) 0.4 ± 22.6 gC m−2 (NPT)]. Our results show, that the N:P stoichiometric imbalance, resulting from N addition (without P), increases the WUE less than the addition of N + P, due to the strong increase in transpiration at NT, which indicates the importance of a balanced N and P content for WUE.